Update README.md

README.md

@@ -27,13 +27,13 @@ base_model: meta-llama/Meta-Llama-3.1-405B-Instruct
   - **Activation quantization:** FP8
 - **Intended Use Cases:** Intended for commercial and research use in multiple languages. Similarly to [Meta-Llama-3.1-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct), this models is intended for assistant-like chat.
 - **Out-of-scope:** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in languages other than English.
-- **Release Date:** 7/24/2024
-- **Version:** 1.0
+- **Release Date:** 8/22/2024
+- **Version:** 1.1
 - **License(s):** [llama3.1](https://huggingface.co/meta-llama/Meta-Llama-3.1-8B/blob/main/LICENSE)
 - **Model Developers:** Neural Magic
 
-Quantized version of [Meta-Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct).
-It achieves an average score of 96.67 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 96.93.
+Quantized version of [Meta-Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct) with the updated 8 kv-heads.
+It achieves an average score of 86.86 on the [OpenLLM](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard) benchmark (version 1), whereas the unquantized model achieves 86.79.
 
 ### Model Optimizations
 
@@ -140,7 +140,7 @@ oneshot(
 
 The model was evaluated on MMLU, ARC-Challenge, GSM-8K, Hellaswag, Winogrande and TruthfulQA.
 Evaluation was conducted using the Neural Magic fork of [lm-evaluation-harness](https://github.com/neuralmagic/lm-evaluation-harness/tree/llama_3.1_instruct) (branch llama_3.1_instruct) and the [vLLM](https://docs.vllm.ai/en/stable/) engine.
-This version of the lm-evaluation-harness includes versions of ARC-Challenge and GSM-8K that match the prompting style of [Meta-Llama-3.1-Instruct-evals](https://huggingface.co/datasets/meta-llama/Meta-Llama-3.1-8B-Instruct-evals).
+This version of the lm-evaluation-harness includes versions of ARC-Challenge, GSM-8K, MMLU, and MMLU-cot that match the prompting style of [Meta-Llama-3.1-Instruct-evals](https://huggingface.co/datasets/meta-llama/Meta-Llama-3.1-8B-Instruct-evals).
 
 ### Accuracy
 
@@ -159,41 +159,51 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge and
   <tr>
    <td>MMLU (5-shot)
    </td>
-   <td>86.25
+   <td>87.41
    </td>
-   <td>86.17
+   <td>87.46
    </td>
-   <td>99.91%
+   <td>100.0%
    </td>
   </tr>
   <tr>
-   <td>ARC Challenge (0-shot)
+   <td>MMLU-cot (0-shot)
    </td>
-   <td>96.93
+   <td>88.11
    </td>
-   <td>96.67
+   <td>88.11
    </td>
-   <td>99.73%
+   <td>100.0%
    </td>
   </tr>
   <tr>
-   <td>GSM-8K (CoT, 8-shot, strict-match)
+   <td>ARC Challenge (0-shot)
+   </td>
+   <td>94.97
    </td>
-   <td>96.44
+   <td>94.97
+   </td>
+   <td>100.0%
+   </td>
+  </tr>
+  <tr>
+   <td>GSM-8K-cot (8-shot, strict-match)
    </td>
    <td>95.98
    </td>
-   <td>99.52%
+   <td>95.75
+   </td>
+   <td>99.76%
    </td>
   </tr>
   <tr>
    <td>Hellaswag (10-shot)
    </td>
-   <td>88.33
+   <td>88.54
    </td>
-   <td>88.34
+   <td>88.45
    </td>
-   <td>100.0%
+   <td>99.90%
    </td>
   </tr>
   <tr>
@@ -201,29 +211,29 @@ This version of the lm-evaluation-harness includes versions of ARC-Challenge and
    </td>
    <td>87.21
    </td>
-   <td>87.45
+   <td>88.00
    </td>
-   <td>100.2%
+   <td>100.9%
    </td>
   </tr>
   <tr>
    <td>TruthfulQA (0-shot, mc2)
    </td>
-   <td>64.64
+   <td>65.31
    </td>
-   <td>64.71
+   <td>65.25
    </td>
-   <td>100.1%
+   <td>99.91%
    </td>
   </tr>
   <tr>
    <td><strong>Average</strong>
    </td>
-   <td><strong>86.63</strong>
+   <td><strong>86.79</strong>
    </td>
-   <td><strong>86.55</strong>
+   <td><strong>86.60</strong>
    </td>
-   <td><strong>99.91%</strong>
+   <td><strong>99.74%</strong>
    </td>
   </tr>
 </table>
@@ -237,12 +247,25 @@ The results were obtained using the following commands:
 ```
 lm_eval \
   --model vllm \
-  --model_args pretrained="neuralmagic/Meta-Llama-3.1-405B-Instruct-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=4096,tensor_parallel_size=8 \
-  --tasks mmlu \
+  --model_args pretrained="neuralmagic/Meta-Llama-3.1-405B-Instruct-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=4096,max_gen_toks=10,tensor_parallel_size=8 \
+  --tasks mmlu_llama_3.1_instruct \
+  --apply_chat_template \
+  --fewshot_as_multiturn \
   --num_fewshot 5 \
   --batch_size auto
 ```
 
+#### MMLU-cot
+```
+lm_eval \
+  --model vllm \
+  --model_args pretrained="neuralmagic/Meta-Llama-3.1-405B-Instruct-FP8-dynamic",dtype=auto,add_bos_token=True,max_model_len=4096,max_gen_toks=1024,tensor_parallel_size=8 \
+  --tasks mmlu_cot_0shot_llama_3.1_instruct \
+  --apply_chat_template \
+  --num_fewshot 0 \
+  --batch_size auto
+```
+
 #### ARC-Challenge
 ```
 lm_eval \